The Futile Cycle

I’m a big fan of Martin Nowak, so it’s great to see his great stuff being proselytized to the masses by the New York Times (found via The Loom). Everyone even remotely interested in evolutionary biology absolutely needs to read Evolutionary Dynamics. Be patient with the math (some calculus, a little linear algebra), because really, it’s absolutely necessary to truly understand the insight he has on the topics. Just read it. It’s got great side chapters on the evolution of linguistic universal grammar, virology, and wonderful stuff on game theory in evolution.

Via Medpundit, an article in Scientific American arguing against race-based medicine. I spoke about this on an aside in a post a few days ago, but now I feel like I have to respond to the Scientific American article. Note that I’m not defending BiDil here, but the concept of race-based medicine in general.

I don’t think the trend for race or ethnicity targeted drugs is a bad thing, overall, because first, we wouldn’t have a shortage of new drugs for everyone, since drug companies would only try for subpopulation-targeted drugs in a case where the drug has failed in a larger population. They want profits; they’re not going to limit the scope of any drug to a subpopulation — race-based or genetics-based or whatever — unless they had no choice; just look at how much they try to expand the indications of already approved drugs. And research produces so many drugs that fail clinical trials anyway that it might as well behoove industry to try drugs on sub-populations, since they’ve brought the drugs so far through the clinical process. So nothing to worry about on the economic side of things.

Secondly, taking race into account is better than not doing so, I think. It is simply wrong and counter-productive to claim that people can’t be grouped into ethnicity based on genetic polymorphisms. There are strong sub-population differences on polymorphisms of drug metabolizing enzymes, so what is toxic to one group may not be to the other. And saying that only 10 drugs or so have evidence to back up race-based indications doesn’t mean that we won’t see more in the future, now that more biological understanding is coming out about such things. But until personal genome sequencing and profiling becomes a lot less expensive and a lot faster than it is now, ethnicity, or even race, is a better substitute than nothing. After all, humans have, on a population scale, very homogeneous genetics, and the variation from continent to continent is pretty clear, with minor exceptions, so ethnicity is a valid and easily detectable, though crude, genetic marker. (Ironically, “black”, which is what BiDil was approved for, is probably the crudest and most meaningless sub-population indicator, since Africa has the most diverse genetic variation in the world.)

I do, however, hope that the FDA does not go lax with its requirements of efficacy. After all, race-targeted or not, the drug still has to work. I don’t believe that a head-to-head of one race versus another would be strictly necessary before approval of a drug, since that brings in ethical issues (testing a drug in populations where researchers expect the drug to fail). Anyway, if a drug really is effective in a larger population, post-approval clinical trials would be a lot easier to test, as they would already have gone through the three previous trials for efficacy and toxicity, and so it would be advantageous to allow the drug company to seek approval based on a limited subset first.

Basically, I see race-based or ethnicity-based medicine as part of the whole “personalized medicine” endeavor. Even though it’s crude, it’s still the first step, and it’s what we have to work with right now. We have drugs targeted based on age and sex already, and biologists very well know that age from birth is just a surrogate for “real” aging, and sex is just a surrogate for “real” hormonal levels, developmental history, and so on. Until we have better biological markers, age, sex, and even race and ethnicity, are good starting points for tailoring medicine to each person. We already do it, so why not with pharmaceuticals, too?

UPDATE: A nice New York Times article on natural selection on human subpopulations.

This brings back many memories of the good old Macintoshes. I found an excellent webpage that explained how to set up an emulation of Dark Castle, with all the appropriate links to files, and now I’m working my way through “Fireball 4. ” I remember playing this on my dad’s old Mac way back in the 1980s on a black-and-white System 6 machine. I don’t know what happened to that old Mac or the Dark Castle disks (probably relegated to archival floppy hell), but now I have my own working copy!

For those of you who want to play along, all you really need is to follow the instructions on that website, and download the appropriate Mini vMac emulator for your system (works for almost any system). The hardest part is getting Dark Castle onto a disk image, and that unfortunately requires borrowing a non-Intel mac for a few minutes. But to relive the past…it’s completely worth it!

If you want other great old-timey Macintosh games, try Macintosh Garden. (All these on Mac Garden require the Classic environment, which means no Intel Macs, sorry. Maybe you can get them to work under emulation, I don’t know.)

EDIT: I got Beyond Dark Castle working as well.

After writing about a series coming to an end in my last book post, I thought I would write about a series (or rather, trilogy) that I’m in the middle of right now.

I discovered Brandon Sanderson on Chad Orzel’s blog review of Elantris, actually. I read the review, though, “eh, maybe I’ll peak at it at some point,” and then saw that Sanderson (the author of Elantris) had actually responded to Chad. At the time, I guess it had never consciously occured to me that authors would go looking for reviews of their own work. Of course, it’s obvious in retrospect that they would, but I didn’t realize at the time how easy it was for them to find such things.

Anyway, Sanderson sort of suggested Mistborn as something that would redeem some of the “saggy” qualities of Elantris (read the post), and — intrigued as much by the author as by the review of Elantris and the suggestion of Mistborn — I found some Mistborn sample chapters on the author’s website. I was hooked, so I checked out the book from the library.

And I liked it! It was engrossing, fun, and I think mostly well-paced. Some of the castle intrigue was a bit long, but there was enough tension that the scenes held my attention. The twists and turns of the book were quite good, and the ending was quite satisfying, leaving the ominous threat looming overhead.

One thing that always plagues fantasy books, especially heavily plotted ones where there are twists and turns and jump surprises (see Harry Potter), is that there are times when one might go, “Woah, where did that come from?” It sucks, because then the suspension of disbelief breaks and you feel the heavy hand of the author. Thankfully, that phenomenon only really struck once or so in the book (near the end; I’ve ROT13′d the spoiler: jura Znefu pbzrf onpx, erfheerpgrq bhg bs abjurer), but strike it did, nonetheless. So not a perfect book, by any means.

But I still liked it very much. The world is very interesting and vivid, which I like. Sanderson manages to sketch the world quite well without feeling as though he’s spending pages on boring exposition and scene-painting. He uses broad strokes to set the scene, and then uses details introduced by the characters in order to fill in the smaller features. The magic system is quite cool. And the characters are mostly very well done, though the main character’s inner monologue is exasperating at times. I’m not a fan of inner monologue, as it’s usually a very blunt technique, but I tolerate it if I like the book otherwise, and I very, very much do.

All in all, a good, very modern, fantasy that hints at greater things to come.

For some reason, today I got to thinking about Bill Clinton’s class day speech at my graduation, where he kept repeating that humans are essentially 99.9% genetically identical; he used it to demonstrate why we shouldn’t emphasize our differences, but instead our similarities and find our connections. It’s an admirable moral goal, anyway, from a political point of view.

But if you look around, it’s relatively easy to find patterns that can roughly categorize people into races. I am Asian. It’s very easy to tell, from my hair color, my skin color, and the bone structure of my face. I’m a bit of a lightweight when it comes to alcohol, and I get that “Asian glow” after a drink or two. Some of my friends are very, very white. Some are Hispanic, some are African-American, some are Middle-Eastern. We can see all this polymorphism, so clearly the 0.1% does make a very big biological difference. After all, genetically we are something like 98% similar to chimpanzees, and yet we see huge differences. 0.1% of 3 billion bases is still 3 million bases.

So I looked around on the internet, and found this old, but wonderful, PLoS interview with Neil Risch, a population geneticist, who in 2002 published a refutation of an NEJM opinion article dismissing the clinical relevance of race. So really, that whole 99.9% thing is a bit misleading, since that 0.1% matters a great deal. So why mention genetics at all? Genetics has no real relevance for big-picture global issues of politics, war and peace, and all that, or for social issues on the level of being kind to each other.

And really, there’s no need to make that sort of weak genetic argument. You can look around and see for yourself that humans are pretty much similar. Everyone is born on basically the same pattern, with mothers and fathers, relatives, with mostly four limbs, two eyes, a nose and mouth, with a brain that loves, a mouth that talks, a heart that pumps, hands that grasp. A smile is a smile in every culture, a laugh is a laugh. There are variations on this theme, especially with modern medicine and technology, but mostly people are the same. We don’t need the human genome project to tell us that. Basing your ethical tenets on really shallow inferences from biology is usually not a good idea. Leave biology for biology and medicine, and just look around.

I borrowed my brother’s copy after he was done, and after a straight 5 hours or so of reading, I’m done with Harry Potter. It took me a bit of time to surface from the deluge, so I’m still pondering whether I really like the book or not.

First of all, it isn’t great. Not masterful, not horrifically tragic, not stupendously inspiring. My first thought was, “it’s finally just done.” I mean, it’s been at least 10 years that I’ve known about the phenomenon, so after all this time, the beginning feels simultaneously long ago (I barely remember even reading the first book) and yet very near. My second thought: “It was ok,” in a “that meal was filling and tasted fine” sense. Perhaps I’m a sentimentalist at heart, but I rather like (mostly) happy endings. My third set of thoughts was more complicated. They were very similar to this review in the Atlantic (there are spoilers after the first paragraph or so). “A children’s story after all.”

In the end, I think I settled somewhere between “I’m glad it’s over” and “I guess I enjoyed parts of it.”

My favorite part about Harry Potter was probably that the world J. K. Rowling constructed was so accessible. Tolkein’s works, though masterful, always had that academic, stuffy, history-tome feel that just made me sleep (I never did get through the Silmarillion). Not only that, but the magic never did have a sense of logic; it was foreign, completely enigmatic. I think that was the point, because then users of magic, like Gandalf, became mythic and inscrutable, a little like Merlin from the Arthurian legends. Phillip Pullman’s world in His Dark Materials was difficult to understand as well, and very creepy. There was just something wrong with it, I can’t really place my finger on it. I think that was intentional, too, in its construction, as the plot was pretty dark and creepy, and from the perspective of the young protagonists, confusing and foreign.

On the other hand, Potterworld was quite charming and friendly, with magic a part of everyday life. There was a degree of casualness that very few other settings had. Magic was common, in a sense, and that made for some very entertaining little details in the world. It wasn’t a rigorous, self-consistent world-system, but just a cartoon of what it would be like to live with magic all around. This friendliness was all, no doubt, designed because the first book or two were meant to make the magical world as inviting as possible to the young audience.

Unfortunately, very little of that was given in the last book, which was a bit disappointing. It was mostly action, with bits of plot-ish-ness in the in-betweens. And, of course, characters die off. It felt formulaic, at times; “oh, she’s offed another one.” Plotting is not J. K. Rowling’s greatest of abilities. But no matter. In the end, the story wrapped up, much in the ways I expected, and then that is that. The series is over. And I’m not unhappy with the way she ended it (even the epilogue, to cap off the whole book of misdeeds, isn’t really that bad).

Now all that has to happen before the second coming is that Robert Jordan actually finish his series. Which may never happen.

I occasionally come across pages like “The Beauty of LaTeX” or other tutorial and article that evangelizes the semantic, archival, aesthetic, and typographical value of LaTeX. Whenever I do, I’m reminded of this reason I once heard from a graduate student:

My advisor really likes to crushingly edit my work; no phrase goes untouched. So when I wrote my thesis, I did it in LaTeX, and since there isn’t a “Track Changes” like in Microsoft Word, he has to edit it by hand, on paper. Suddenly, only really relevant suggestions and edits were given back to me.

(Of course, LaTeX does have a sort of “track changes” type package, but it’s not quite as easy to use as MS Word, and requires knowledge of LaTeX.)

For those of you in the Cambridge, MA area, seriously, one of the best Italian restaurants I know of is Basta Pasta Trattoria. My friend introduced me to the restaurant in the very last week of college, but fortunately my family lives around Boston, so I’ve been able to return. Very great (I’d give it a 24 on the Zagat scale) but also insanely cheap. I mean, a huge bowl of delicious, heavenly risotto (try the white risotto with thyme, lemon, prosciutto, and goat cheese) for less than $9? It’s like a liter and a half of the stuff, too. What’s better for a poor student/post-doc? Also, the mini arancinis (fried rice balls and fontina cheese, with tomato sauce) are awesome.The only problem is that they’re a bit of a walk away from any T station (basically at the intersection of Putnam Ave and Western), and it’s in a residential neighborhood (i.e. little to no parking) but they do catering. I’m not sure if they do delivery. I go there by walking from Harvard Square, which is like 15 mins. The walk there is easy, the walk bad is hard because I’m usually stuffed full.

Part of the problem with looking at a field like computational biology or systems biology (which I have some interest in) is that it’s hard to figure out how to narrow the field into one project that I can do for graduate school.

In synthetic organic chemistry, this was a relatively easy proposition. Pick a molecule, not too big, not too small, with interesting complexity, and make it. Then you’re done with your Ph.D. (course, this is much easier said than done). In traditional biochemistry, molecular biology, or cell biology, you’d focus on one protein, one pathway, one set of activities, one phenotype, and then figure out what they do, how they interact, what affects them, etc. When you generate enough publications and have a good, detailed story to tell about the systems, you’re done.

But computational biology feels like a whole ‘nother thing. It’s probably because I’m still not quite as familiar with that field, but it seems very methodological, more than anything else, and so it’s hard to find a particular angle to break down the problem and get a story out of it for a thesis. There’s not just one specific model system to work on, or one biochemical system that one can mine. At some point, it feels a little arbitrary, that you choose a system hoping that your hammer will find a nail you can hit; it could be that you choose wrong and just get screwed (I couldn’t help myself with that pun).

I remember reading about Eugene Shakhnovich’s work on protein folding, and he always had a particular angle to hit the topic with: his lattice folded proteins. One of the professors that I talked to at my new university tends to focus on bacterial chemotaxis. I need a good focus point by which to approach the field. Large-scale topics for data-mining, such as the ENCODE project, just seem too big for me to do as part of my Ph.D. project, and it would probably make my Ph.D. drag on and on in a way that would both make me depressed and make my work harder.

Perhaps I can find a particular algorithm to work on, if I decide to do a “big biology” sort of project. A particular technique, perhaps? I feel like doing a particular methodology is the only manageable way to approach “looking at systems-level behavior.” Otherwise, you’re relying on luck (pick a system and hope you can use your hammer). Because systems biology isn’t really a field, right? It’s just a methodology.

Maybe that’s why people like the reductionist mode of research so much? At least it’s easy to find a story to tell. I guess it’s a little like history. Everyone wants a handle, some sort of focal point. So you get focal points, like Alexander the Great, or Napolean. Even when studying social trends and situations, such as poverty and crime, you get the Italian Mafia, Tammany Hall, Five Points. Focal points.

So here I am, trying to wade through the literature and find a good focus.